Gunshot sound effects you generate yourself

A gunshot is two sounds: a very short, very loud report, and the environment's response to it. The report tells you what the weapon is; the response tells you where you are standing. Most unconvincing gunshots in an edit have a plausible report and no environment, so they sound like samples. These prompts always name the space, because it is at least half of what the audience hears.

Listen: 10 samples from the gunshots collection

Every clip shows the exact prompt that produced it. The render script reads its prompts from the same file as this page. Copy one, change a word and generate your own.

  • Pistol indoors

    via audiocpp-stable-audio-3-sfx

    9mm pistol fired once inside a concrete stairwell, sharp cracking report with hard slapping echoes, close

  • Pistol outdoors

    via audiocpp-stable-audio-3-sfx

    pistol fired once in an open field, sharp report with a fast dry crack and almost no reflection, distant treeline

  • Distant rifle

    via audiocpp-stable-audio-3-sfx

    hunting rifle fired far away across a valley, delayed crack with a long rolling echo between hillsides

  • Shotgun and pump

    via audiocpp-stable-audio-3-sfx

    pump-action shotgun fired then racked, heavy blast with a shell ejecting onto concrete, alley reverb

  • Automatic burst

    via audiocpp-stable-audio-3-sfx

    automatic rifle firing a short burst in a city street, rapid overlapping reports with building echoes

  • Suppressed shot

    via audiocpp-stable-audio-3-sfx

    suppressed pistol shot, muted mechanical thump with the slide cycling clearly audible, close and dry

  • Revolver

    via audiocpp-stable-audio-3-sfx

    revolver fired once, deep booming report with a cylinder ring, open outdoor space with a rolling tail

  • Magazine reload

    via audiocpp-stable-audio-3-sfx

    pistol magazine released and dropped, a fresh one inserted and the slide released, very close mic, dry

  • Ricochet

    via audiocpp-stable-audio-3-sfx

    bullet ricocheting off metal, sharp impact with a whining pitched departure, industrial space, echoing

  • Distant firefight

    via audiocpp-stable-audio-3-sfx

    distant firefight heard from several streets away, scattered muffled reports with long urban echoes

Prompts that work

These models respond to physical description. Name the material, the action and the space.

  • Indoors

    9mm pistol fired once inside a concrete stairwell, sharp cracking report with hard slapping echoes, close

    Hard enclosed spaces produce discrete slap-back echoes instead of smooth reverb, and naming the material gets you that reflection.

  • Outdoors

    pistol fired once in an open field, sharp report with a fast dry crack and almost no reflection, distant treeline

    "Almost no reflection" makes an exterior shot sound exterior, and the treeline adds one distant return so it does not sound lifeless.

  • Distance

    hunting rifle fired far away across a valley, delayed crack with a long rolling echo between hillsides

    At distance the echo becomes the main event and the report becomes small, so prompting the echo prompts the distance.

  • Mechanism, not report

    pistol magazine released and dropped, a fresh one inserted and the slide released, very close mic, dry

    Weapon handling is used far more often than firing and is almost never in stock packs at usable length.

Making gunshots sounds that fit

Always name the environment

A gunshot's report lasts a few milliseconds; the rest is the room, the street or the valley responding. That is why the same weapon is unrecognisable indoors and outdoors, and why a dry stock shot never sits in a scene. Make the space a required part of every gunfire prompt. When a shot does not fit the picture, change the space before the weapon.

Generate the handling sounds first

Slides racking, magazines dropping, hammers cocking and shells hitting concrete are quiet, specific and instantly readable. They carry tension because they signal intent, where a shot signals consequence. They are also short and cheap to generate. If you are building a weapon set, make the handling sounds first; scenes are often built entirely from them.

Distance delays and smears the crack

Sound travels roughly a third of a kilometre per second, so a distant shot arrives noticeably after the muzzle flash, and smeared: the sharp transient has been absorbed and what is left is the rolling echo. Prompting "delayed crack" and "long rolling echo" gives you that, and it reads as distance immediately. A quiet close gunshot sounds like a quiet gunshot, not a distant one.

Why generate locally

The models run on your machine, which changes the cost, the privacy and what you can automate.

Offline after download

Weights are fetched once. After that, generation, the library and the API all work with the network off. Everything lives in one SQLite file and one folder of WAVs under ~/.soundbox, and nothing is uploaded.

Your agent can use it

Soundbox runs an MCP server in the same process, with tools for generating and listing sounds. Generated files come back as absolute paths, so an agent can copy them into a project.

Licences stated up front

Every model's terms are shown before you download it, and non-commercial weights need an explicit confirmation. Soundbox is MIT. The licence of the model that made a sound governs what you may do with it.

Connect a coding agent
claude mcp add --transport http soundbox http://127.0.0.1:17494/mcp
Generate over HTTP
curl -X POST http://127.0.0.1:17494/api/generate \
  -H 'Content-Type: application/json' \
  -d '{"model":"audiocpp-stable-audio-3-sfx","prompt":"8-bit coin pickup blip","duration_s":1}'

Frequently asked questions

Will these pass for real firearm recordings?
For most game and video work, yes, especially the handling sounds and anything at distance, where the environment dominates. Close, dry, high-calibre reports are the hardest case for any generative model: the transient is extremely short and extremely loud, and an approximation has nowhere to hide. If you need that sound at feature-film scrutiny, layer a generated tail under a recorded crack. For everything else, a generated shot in a well-described space holds up.
Can I use these sounds in a commercial video or game?
It depends on the model that made them. Soundbox shows each model's licence before you download its weights. The clips on this page were rendered with Stable Audio 3.0 Small SFX, which uses the Stability AI Community License: commercial use is free below USD $1M annual revenue, and above that Stability requires an enterprise licence. Soundbox itself is MIT, but the audio is governed by the licence of the model that made it. Read that licence before you ship; this summary is not legal advice.
Does this work offline?
Yes, once the weights are on disk. Generation, the library and the HTTP API all run locally against one SQLite file and one folder of WAVs under ~/.soundbox. There is no account and no API key, and nothing is uploaded. The only step that uses the network is the initial model download.
What hardware does it need?
The clips here were rendered on a consumer NVIDIA GPU at roughly ten seconds of wall time per clip, through audio.cpp's quantised GGUF runtime, which needs little VRAM. CPU-only generation is supported, but expect minutes per clip instead of seconds: fine for a handful of sounds, slow for a whole library.
Can an agent generate these for me?
Yes. While the app is open, Soundbox runs an MCP server at http://127.0.0.1:17494/mcp with tools for generating, listing and mixing sounds. Every generated file comes back as an absolute path, so a coding agent can render the sound a scene needs and put it straight into the project. The several hundred clips across these pages were made this way.